New software can detect COVID-19 outbreaks in U.S. counties

New tool can detect COVID-19 outbreaks in U.S. counties

Credit: Unsplash / CC0 Public Domain

A new online machine learning-based tool developed by researchers at Massachusetts General Hospital (MGH), Harvard Medical School (HMS), Georgia Tech, and Boston Medical Center enables early detection of COVID-19 Outbreaks in various US states. The COVID-19 outbreak detection tool is updated two to three times a week and predicts how quickly an outbreak will spread in a given county by estimating the doubling time of COVID-19 cases.

To make these predictions, the tool takes into account reported COVID-19 cases and deaths, face mask mandates, social distancing guidelines, changes in tests performed, positive test rates, and the Centers for Disease Control and Prevention's Social Vulnerability Index (which the Health assesses) -related resilience of individual communities to external pressures such as natural disasters or man-made disasters or disease outbreaks).

The tool provides an interactive map and "data explorer" that allows users to select a specific county to view the population of that county, the total number of new cases of COVID-19 in the last week, the average daily cases in the last week and to display the COVID. 19 Doubling rate (ie how many days it takes for the number of cases in a given county to double).

"While previous mitigation efforts have focused on state-level measures – for example, an entire state lockdown – by exposing local outbreaks, policymakers can implement county-level measures, such as closing restaurants in a single county, to effectively contain the pandemic," said Dr. Jagpreet Chhatwal, Associate Director at MGH Institute for Technology Assessment and Assistant Professor at HMS. "The model is able to detect many of the previous outbreaks within a few days."

"To effectively combat the pandemic, it is important to detect an outbreak in good time so that the affected area can be isolated and the spread of COVID-19 infections to neighboring areas can be minimized. However, this may be possible for a number of reasons." It takes days or even weeks for people to manually detect an outbreak. Our data-driven machine learning solution accelerates and automates this process significantly, "said Turgay Ayer, Ph.D., director of business intelligence and healthcare analytics at the Center for Health and Humanitarian Systems and associate professor at Georgia Tech.

Using the COVID-19 Outbreak Tool, the research team checked an outbreak in Johnson County, Iowa last week that has been linked to an outbreak at the University of Iowa. The tool also identified several districts where outbreaks could now occur. These include Harrisonburg County in Virginia, Wheeler County in Georgia, Monroe County in Indiana, and Whitman County in Washington, where infections double in less than a week.

Follow the latest news on the Coronavirus (COVID-19) outbreak

More information:
The COVID-19 outbreak detection tool:

Provided by
Massachusetts General Hospital

New Tool Can Detect COVID-19 Outbreaks in U.S. States (2020, Sep 16)
accessed on September 16, 2020

This document is subject to copyright. Apart from fair treatment for the purpose of private study or research, no
Part may be reproduced without written permission. The content is provided for informational purposes only.


Please enter your comment!
Please enter your name here